After binging on “House of Cards” and “Orange Is The New Black,” how does Netflix determine what a user should watch next? Netflix has a complex recommendation algorithm that is worked on by 800 engineers at Netflix’s Silicon Valley officers so users can get more viewing ideas.
Until Netflix rolled out its user profiles recently, it used the entire watching habits of a household to make its recommendations. This means that one person who enjoys watching “Star Trek” and “Doctor Who” will also get recommendations to watch “Glee” and “Once Upon a Time” because another person in the house watches that content. Now that they’ve added user profiles, the recommendations focus on the individual’s viewing tastes.
With new features that help improve viewer recommendations, Wired recently spoke to two of Netflix’s engineers who talked about what they consider when determining what to show users as recommendations. Engineering director Xavier Amatriain told Wired that Netflix knows what users watched, rated and rated based on the time, date and their device.
He added: “We even track user interactions such as browsing or scrolling behavior. All that data is fed into several algorithms, each optimized for a different purpose. In a broad sense, most of our algorithms are based on the assumption that similar viewing patterns represent similar user tastes. We can use the behavior of similar users to infer your preferences.”
As far as predetermined ratings, Netflix’s Vice-President of Product Innovation and Personalization Algorithms Carlos Gomez-Uribe said to Wired: “Testing has shown that the predicted ratings aren’t actually super-useful, while what you’re actually playing is. We’re going from focusing exclusively on ratings and rating predictions to depending on a more complex ecosystem of algorithms.”
Amatriain said that Netflix is working to add context to its recommendations, noting that it has the data based on what users watch, when they watch it and how they watch it.
He added: “But implementing contextual recommendations has practical challenges that we are currently working on. We hope to be using it in the near future.”
You may also like: